Bacteriopurpurinimides: Highly stable and potent photosensitizers for photodynamic therapy

Article abstract of DOI:10.1021/jm010400c

The in situ conversion of the unstable bacteriochlorophyll a present in Rhodobacter sphaeroides produced highly stable bacteriopurpurin-18 which in a sequence of reactions was converted into a series of alkyl ether analogues of bacteriopurpurin-18-N-alkylimides with long wavelength absorption near 800 nm. The effective photosensitizers were found to localize in mitochondria but did not show any specific displacement of ³H-PK11195, suggesting that the mitochondrial peripheral benzodiazepine receptor is not the cellular binding site for this class of compounds.

Full text of DOI:10.1021/jm010400c

J . Med. Chem. 2002, 45, 255-258
255
is why most of the sensitizers currently under clinical
evaluation for PDT are porphyrins or porphyrin-based
molecules.
Ba cter iop u r p u r in im id es: High ly Sta ble
a n d P oten t P h otosen sitizer s for
P h otod yn a m ic Th er a p y
It is well established that both absorption and scat-
tering of light by tissue increases as the wavelength
decreases and that most efficient sensitizers are those
which have strong absorption bands from 700 to 800
nm. Light transmission by tissues drops rapidly below
550 nm; however, it doubles from 550 to 630 nm and
doubles again from 630 to 700 nm. This is followed by
an additional 10% increase in tissue penetration as the
wavelength increases toward 800 nm.2 Another reason
to set the ideal wavelength for PDT at 700-800 nm is
due to the availability of easy to use diode lasers.
Although diode lasers are now available at 630 nm
(where clinically approved Photofrin absorbs), photo-
sensitizers with absorptions between 700 and 800 nm
in conjunction with diode lasers are still desirable for
treating deeply seated tumors.
Most of the naturally occurring bacteriochlorins have
absorptions between 760 and 780 nm, and several have
been studied by various investigators3 for their use as
photosensitizers for PDT. They were found to be ex-
tremely sensitive to oxidation, resulting in a rapid
transformation into the chlorin state which generally
has an absorption maxima at or below 660 nm. Fur-
thermore, if a laser is used to excite the bacteriochlorin
in vivo, oxidation may result in the formation of new
chromophore absorbing outside the laser window, re-
ducing the photodynamic efficacy.3 Due to the desirable
photophysical properties and promising in vitro/in vivo
photosensitizing efficacy of bacteriochlorins, there has
been increasing interest in the synthesis of stable
bacteriochlorins either from bacteriochlorophyll a or
from the other related tetrapyrrolic systems.
Resu lts a n d Discu ssion . Ch em istr y: In our previ-
ous studies with various photosensitizers derived from
pyropheo-phorbide a, chlorin-e6, benzoporphyrin deriva-
tives, and N-alkylpurpurinimides,3 lipophilicity has
proven to be an important molecular descriptor that
often was found to be well-correlated with the bioactivity
of drugs. In these analogues, the overall lipophilicity of
the molecule was altered by introducing a series of the
alkyl ether side chains with variable carbon units. In
continuation of our studies to investigate the effect of
such substituents in longer wavelength absorbing com-
pounds related to bacteriochlorins, we developed a
simple and efficient methodology for the preparation of
stable bacteriochlorins.
It has been shown that certain chlorophyll a ana-
logues containing a five-member isocyclic ring (λmax 660
nm) can be converted into the related chlorins bearing
either a fused six-member anhydride or imide ring
systems and named as purpurin-18 and purpurinimide,
respectively. The presence of such fused ring systems
extended their long wavelength absorptions from 660
to 700 nm.3 Despite their high stability in vitro, results
obtained from in vivo reflectance spectroscopy indicated
that compounds containing the six-member imide ring
system were more stable in vivo than the photosensi-
tizers with anhydride and isoimide ring systems.3
Yihui Chen,† Andrew Graham,‡ William Potter,†
J anet Morgan,‡ Lurine Vaughan,† David A. Bellnier,†
Barbara W. Henderson,† Allan Oseroff,‡
Thomas J . Dougherty,† and Ravindra K. Pandey*,†
Photodynamic Therapy Center and Department of
Dermatology, Roswell Park Cancer Institute,
Buffalo, New York 14263
Received August 23, 2001
Abstr a ct: The in situ conversion of the unstable bacterio-
chlorophyll a present in Rhodobacter sphaeroides produced
highly stable bacteriopurpurin-18 which in a sequence of
reactions was converted into a series of alkyl ether analogues
of bacteriopurpurin-18-N-alkylimides with long wavelength
absorption near 800 nm. The effective photosensitizers were
found to localize in mitochondria but did not show any specific
3
displacement of H-PK11195, suggesting that the mitochon-
drial peripheral benzodiazepine receptor is not the cellular
binding site for this class of compounds.
In tr od u ction . The porphyrins and related tetrapyr-
rolic systems are among the most widely studied of all
macrocyclic compounds.1 In fact, in one capacity or
another these versatile molecules have influenced nearly
all disciplines in chemistry. Porphyrins are 18π-electron
aromatic macrocycles that exhibit characteristic optical
spectra with a strong π-π* transition around 400 nm
(Soret band) and usually four Q bands in the visible
region. In the porphyrin system, two of the peripheral
double bonds in opposite pyrrolic rings are cross-
conjugated and are not required to maintain aromatic-
ity. Thus reduction of one or both of these cross-
conjugated double bonds (to give chlorins and bacterio-
chlorins, respectively) maintains much of the aroma-
ticity, but the change in symmetry results in batho-
chromically shifted Q bands with high extinction coef-
ficients. Nature uses these optical properties of the
reduced porphyrins to harvest solar energy for photo-
synthesis with chlorophylls and bacteriochlorophylls as
both antenna and reaction-center pigments.1 The long
wavelength absorption of these natural chromophores
led to explorations of their use as photosensitizers in
photodynamic therapy (PDT).2
PDT is based on the interaction of a photosensitizer
retained in tumors with photons of visible light, result-
ing in the formation of singlet oxygen (1O2), the putative
lethal agent.2 To achieve an effective destruction of
tumor cells, a high quantum yield of singlet oxygen is
required. Even in the absence of heavy atom substitu-
tion(s) and coordination of transition-metal ions, por-
phyrin systems generally satisfy these criteria, and that
* Address for correspondence: Ravindra K. Pandey, Ph.D., PDT
Center, Roswell Park Cancer Institute, Buffalo, NY 14263. Phone:
(716) 845-3203. Fax: (716) 845-8920. E-mail: ravindra.pandey@
roswellpark.org.
† Photodynamic Therapy Center.
‡ Department of Dermatology.
10.1021/jm010400c CCC: $22.00 © 2002 American Chemical Society
Published on Web 12/18/2001

256 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 2
Letters
Sch em e 1. Synthesis of Various Alkyl Ether Analogues
of Bacteriopurpurin-18-N-hexylimidea
F igu r e 1. In vitro PDT efficacy of bacteriochlorin 4-10 in
RIF tumor cells at a dose of 1.5 µM at 48 h MTT. The cells
were treated with light (780 nm) after 4 h incubation (see the
text). Control: cell exposed to light without any photosensi-
tizer.
into mice at a dose of 5.0 µmol/kg. All photosensitizers
showed a greater uptake in tumor than skin at 24 h
postinjection, indicating some degree of tumor selectiv-
ity, the ratio varying less than 2.5 (compound 5) to
nearly 5 (compound 4).
(b) In Vitr o P h otosen sitizin g Effica cy. RIF cells
were grown in R-DMEM with 10% fetal calf serum,
penicillin, and streptomycin.2 Cells were maintained in
5% CO2 and 95% air at 100% humidity. For phototox-
icity studies, RIF cells were plated in 96-well plates at
a density of 1 × 104 cells per well, in complete media.
Twenty-four hours later compounds were added at a
final concentration of 1.5 µM. After a 3 h incubation in
the dark at 37 °C, the cells were irradiated with 780
nm laser light from an argon pumped dye laser using
fluences of 0-10 J /cm2 at a dose rate of 5.6 mW/cm2.
After PDT, the cells were washed once and placed in
complete media and incubated for 48 h. Then 10 µL of
4 mg/mL solution of MTT was added to each well. After
4 h incubation at 37 °C, the MTT + media were removed
and 100 µL of DMSO was added to solubilize the
formazin crystals. The 96-well plate was read on a
microtiter plate reader at an absorbance of 560 nm. The
results were plotted as percent survival of the corre-
sponding dark (drug no light) control for each compound
tested. Each data point represents the mean from three
separate experiments, and the error bars are the
standard deviation. Each experiment was done with five
replicate wells. The results are summarized in Figure
1, and it can be seen that all bacteriochlorins (4-10)
produced a significant photosensitizing activity.
(c) In Vivo P h otosen sitizin g Effica cy. C3H mice
(6 mice/group) were subcutaneously implanted with
radiation induced fibrosarcoma (RIF) cells.2 The tumors
were allowed to grow for a period of 5 days to a size of
about 4-5 mm. In our preliminary studies, the treat-
ment consisted of a tail vein injection of photosensitizer
7 at a dose of 0.2 µmol/kg. These mice were exposed to
light for 30 min (135 J /cm2) at λmax 785 nm (in vivo
absorption) at 24 h postinjection, and the tumor re-
sponse (reappearance of tumor) was recorded daily. The
other bacteriochlorins 4-6 and 8-10 were treated
under similar conditions. Compared to in vitro results
that showed a similar PDT efficacy of all bacteriochlo-
a
Reagents: (a) KOH/1-propanol, H2SO4; (b) H2N(CH2)5CH3; (c)
CH2N2; (d) KOH/MeOH; (e) NaBH4; (f) o-dichlorobenzene (140 °C,
2 min); (g) HBr/AcOH; (h) R2OH; (i) HBr gas.
Therefore, for synthesizing stable bacteriopurpurinim-
ides, Rhodobacter sphaeroides enriched with bacterio-
chlorophyll a was reacted with KOH/1-propanol solution
and the air was bubbled through the reaction mixture
for 1.5 h. The solution was acidified to pH ∼3 and
extracted with CH2Cl2/THF, and after heating at 45 °C,
bacteriopurpurin-18 1 was isolated as a solid product.
It was then converted into N-hexylpurpurinimide 4 by
following the methodology depicted in Scheme 1. De-
pending on the nature of substituent present at position-
3, two different approaches (HBr/AcOH or HBr gas
method) were used for introducing the alkyl ether side
chain at position-3 of the macrocycle (Scheme 1).
Compared to the HBr/AcOH method, the HBr gas
procedure was found to be more efficient for the
synthesis of purpurinimides 7-10 with variable lipo-
philicity. These bacteriopurpurinimides exhibit long
wavelength absorption at 780 nm (ꢀ ) 47 300), and if
compared to Photofrin (622 nm, ꢀ ) 5000), HPPH (660
nm, ꢀ ) 45 500), and alkyl ether analogues of purpurin-
N-alkylimide (700 nm, ꢀ ) 45 000), significant red shifts
of 158, 120, and 80, respectively, were observed. The
structures of the newly synthesized compounds were
1
confirmed by H NMR and mass spectrometry.
In Vit r o a n d in Vivo Biologica l St u d ies: (a )
Deter m in a tion of Tu m or Up ta k e by in Vivo Re-
flecta n ce Sp ectr oscop y. The bacteriopurpurinimides
4-10 had limited water solubility. Therefore, for bio-
logical studies, these compounds were formulated in a
1% Tween 80/5% dextrose solution, and the drug
concentrations were determined on the basis of their
extinction coefficient values.
The tumor and skin uptake of each compound was
determined by in vivo reflectance spectroscopy. For
measuring tumor uptake, the compounds were injected

Letters
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 2 257
(d ) Deter m in a tion of In tr a cellu la r Loca liza tion
by F lu or escen ce. Determination of the critical site(s)
of photosensitizer binding has long been a goal of the
PDT research. It has been shown that effective photo-
dynamic agents have very diverse patterns of localiza-
tion, based on structure, lipophilicity, charge, and
amphilicity. Two sites of localization are predominant
in the lysosomes and the mitochondria.4
In our attempt to investigate the site(s) of localization
of bacteriopurpurinimide analogues, photosensitizers
were co-incubated in the RIF and FaDu cell lines with
Rhodamine-123 (Molecular Probes, Eugene, OR) or
Fluorspheres, 0.1 µm diameter latex beads labeled with
fluorescein isothiocyanate 1/10000 dilution (Molecular
Probes, Eugene, OR), which target mitochondria and
lysosomes, respectively. Briefly, 1 × 105 cells were
plated on poly-L-lysine coated coverslips and grown for
2 days. Cells were incubated with 1 µM of the individual
bacteriochlorins 7-10 for 4 or 24 h at 37 °C, 5% CO2 in
the dark, then washed with fresh growth medium for 1
h to remove loosely bound drug, and gently rinsed with
PBS immediately prior to microscopy (Zeiss Axiovert 35,
Carl Zeiss, Inc., Germany). Localization was similar in
both cell lines and at both time points. For the co-
incubation studies, fluorospheres were added 1/10000
dilution of the stock for 24 h, or R-123 at 1.0 µM, for
the final 30 min before washing with PBS and examin-
ing by microscopy. Fluorescence was imaged with a CCD
camera and intensifier as previously described.5 The
images were captured and processed by Metamorph,
version 4.0 (Universal Imaging Corp., Westchester, PA).
The cells were illuminated by a mercury arc lamp
filtered through a filter cube containing a 530-585 nm
excitation filter, a 600 nm dichroic filter, and a 615 nm
long-pass emission filter for detection of photosensitiz-
ers. R-123 and Fluorspheres were detected using a 450-
490 nm excitation filter, a 510 nm dichroic, and a 520-
560 nm long pass emission filter. Images of photosensi-
tizer and Rhodamine-123 or Fluorspheres localization
were taken in rapid succession. The imaging data
clearly indicate that bacteriopurpurinimides 8-10 local-
ize to the same subcellular regions as Rhodamine-123,
suggesting that these compounds selectively localize in
F igu r e 2. In vivo photosensitizing efficacy of bacteriochlorins
7 (-3-), 8 (-O-), 9 (-0-), and 10 (-4-) in C3H mice (6
mice/group) bearing RIF tumors at a dose of 0.2 µmol/kg. The
mice were treated with laser light (780 nm, 135 J /cm2) at 24
h postinjection of the drug. Control (-b-): 12 mice were
exposed to light without any photosensitizer.
rins, a remarkable difference was observed in in vivo
activity. Under similar treatment conditions, bacterio-
purpurinimides containing an acetyl-4, (1′-hydroxy)ethyl
substituent 5, and vinyl group 6 at position-3 did not
show any PDT efficacy. Among the alkyl ether ana-
logues, the 3-(1-methoxy)ethyl 7 (log P ) 8.18) and 3-(1-
propyloxy)ethyl 8 (log P ) 9.16) produced limited
activity. Bacteriochlorins 9 containing heptyl ether
group 9 (log P ) 11.19) showed excellent PDT efficacy
(5/6 mice were tumor free at day 90), whereas the
corresponding bacteriochlorin 10 with a decyl ether
group (log P ) 12.72) was found to be less effective (2/6
mice were tumor free on day 90). The results are
summarized in Figure 2. The limited activity of photo-
sensitizers 4-8, suggest that for a compound to be
biologically effective, besides the tumor uptake, the
clearance of the drugs might play a significant role in
its activity. These findings show that, to some extent,
increasing the length of the alkyl ether carbon chain
increases the photosensitizing capacity; however, it then
diminishes by further increase of the overall lipophil-
icity. To establish a QSAR in this class of compounds,
further detailed pharmacokinetic and pharmacodynamic
studies with these and other related analogues at
variable drug/light doses and time intervals are cur-
rently in progress.
F igu r e 3. Comparative intracellular localization of bacteriopurpurinimides 7 and 9 with Fluorspheres and Rhodamine 123,
respectively.

258 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 2
Letters
Con clu sion . In summary, we present an efficient
approach to in situ conversion of extremely unstable
bacteriochlorophyll a present in bacteria Rb. sphaeroides
into into a series of stable N-hexylimide analogues
exhibiting long wavelength absorption near 780-815
nm. Preliminary in vitro and in vivo results of some of
the related 3-(1′-alkoxy)ethyl- analogues demonstrate
their potential as promising photosensitizers for pho-
todynamic therapy. Among the bacteriopurpurinimides
studied so far, the methyl ether analogue 7 that localizes
in lysosomes produced limited in vivo PDT efficacy,
whereas two of the more effective higher alkyl ether
derivatives (9, 10) were found to localize in mitochon-
dria. The displacement studies with PK11195 suggest
that, unlike the hexyl ether derivative of pyropheophor-
bide a (HPPH) and benzoporphyrin derivatives (BPD),
the effective alkyl ether derivatives of bacteriopurpurin-
N-hexylimides that localize in mitochondria did not have
significant affinity toward the peripheral benzodi-
azepine receptor. Among the compounds tested so far,
the bacteriopurpurinimide 9 (log P ) 11.19) was found
to be most effective at a dose of 0.2 µmol/kg. To select
the best candidate, further in vitro and in vivo studies
with this and other related bacteriochlorins are cur-
rently in progress.
3
F igu r e 4. Displacement of H-PK11195 by PK11195 (-9-)
and with bacteriochlorin 9 (- b-).
mitochondria (see Figure 3, a representative example).
In other experiments, the results obtained by using
Fluorospheres as a counterstain indicated that in
contrast to bacteriopurpurinimides 8-10, 3-(1-meth-
oxy)ethyl purpurinimide 7 that had shown limited in
vivo PDT efficacy was found to localize in lysosomes
which is generally thought of as an inefficient site for
producing photodamage.4
In previous studies it has been implied by us and
others that certain photosensitizers that show mito-
chondrial localization exhibit peripheral benzodiazepine
receptor (PBR) binding which may be an important
target for PDT.4 Therefore, using the peripherally active
benzodiazepine ligand PK11195 as a tool for measuring
binding and displacement at the PBR, we looked for
evidence to support this hypothesis with the bacterio-
purpurinimides. We determined the capacity of biologi-
cally effective bacteriopurpurinimides to displace
PK11195 from its specific cellular binding site, the IQ
site on PBR. Specific radioactive binding assays using
3H-labeled ligand were used to determine the dissocia-
tion constant (Kd) of PK 11195 in whole cells. Initially,
saturation binding of PK11195 was determined on FaDu
and RIF cells in the presence and absence of excess
unlabeled ligands. Bound and free concentrations of
ligand were calculated, and binding affinities were
determined by Scatchard analysis. Nonspecific binding
of PK11195 was 10% or less. The Kd of PK11195 was
23 nM in FaDu and 44 nM in RIF, which agrees with
values quoted in the literature in the low nanomolar
range of 3-50 nM.5a The concentration of 3H-PK11195
which causes 50% of the maximum binding was used
for subsequent displacement studies. Preliminary ex-
periments with increasing concentration of mitochon-
drially localized bacteriopurpurinimide did not indicate
any specific displacement of 3H-PK11195. These results
are in contrast to those obtained from HPPH and benzo-
porphyrin derivatives (BPD)5b and suggest that the PBR
is not the target for the highly effective bacteriopur-
purinimide analogues. The results obtained from the
most effective photosensitizer are depicted in Figure 4.
Ack n ow led gm en t . This investigation was sup-
ported by the NIH (CA 55791) and the shared resources
of the RPCI support grant (P30CA16056). MS analyses
were obtained from Michigan State University, East
Lansing.
Su p p or tin g In for m a tion Ava ila ble: 1H NMR, mass
spectrometry data, electronic absorption spectra, and tumor
vs skin uptake results of new compounds are available free of
charge via the Internet at http://pubs.acs.org.
Refer en ces
(1) The Porphyrin Handbook; Kadish, K. M., Smith, K. M., Guilard,
R., Eds.; Academic Press: San Diego, 2000; Vol. 1-10.
(2) Dougherty, T. J .; Gomer, C.; Henderson, B. W.; J ori, G.; Kessel,
D.; Korbelik, M.; Moan, J .; Peng, Q. Photodynamic Therapy. J .
Natl. Cancer Inst. 1998, 90, 889-900 and references therein.
(3) Pandey, R. K.; Zheng, G. Porphyrins as Photosensitizers in
Photodynamic Therapy. The Porphyrin Handbook; Kadish, K.
M., Smith, K. M., Guilard, R., Eds.; Academic Press: San Diego,
2000; Vol. 6 and references therein.
(4) (a) Kessel, D.; Luo, Y. Mitochondrial photodamage and PDT-
induced apoptosis. Cell Diff. 1999, 6, 28-35. (b) MacDonald, I.
J .; Dougherty, T. J . Basic Principles of photodynamic therapy.
J . Porphyrins Phthalocyanines 2001, 5, 105-129 and references
therein.
(5) (a) Verma, A.; Facchina, S. L.; Hirsch, D. J .; Song, S.-Y.; Dillahey,
L. F.; Williams, J . R.; Snyder, S. H. Photodynamic tumor
therapy: Mitochondrial benzodiazepine receptors as a thera-
peutic target. Mol. Med. 1998, 4, 40-45. (b) Pandey, R. K.;
Morgan, J . Unpublished results.
J M010400C